JPH0199093A - Reference pattern generator for voice recognition - Google Patents
Reference pattern generator for voice recognitionInfo
- Publication number
- JPH0199093A JPH0199093A JP62254731A JP25473187A JPH0199093A JP H0199093 A JPH0199093 A JP H0199093A JP 62254731 A JP62254731 A JP 62254731A JP 25473187 A JP25473187 A JP 25473187A JP H0199093 A JPH0199093 A JP H0199093A
- Authority
- JP
- Japan
- Prior art keywords
- pattern
- reference pattern
- voice
- standard pattern
- input
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000010586 diagram Methods 0.000 description 3
- 238000000034 method Methods 0.000 description 3
- 238000001615 p wave Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 1
Abstract
Description
【発明の詳細な説明】
〔産業上の利用分野〕
本発明は、不特定話者の音声認識等に用いられる標準パ
ターンの作成装置に関するものである。DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a standard pattern creation device used for speech recognition of unspecified speakers.
不特定話者の発声した単語音声を認識するには、多数話
者の各単語音声毎に認識用の標準パターンを作成してお
き、不特定入力音声のパターンと各標準パターンとを比
較し、最も近似した標準パターンの単語を認識結果とす
る手段が用いられておυ、この際、多数話者の各単語音
声から代表的なパターンを選択するには、クラスタリン
グの手法が用いられている。In order to recognize word sounds uttered by unspecified speakers, standard patterns for recognition are created for each word sound of multiple speakers, and the pattern of the unspecified input sound is compared with each standard pattern. A method is used in which the word of the standard pattern that is most similar is used as the recognition result, and in this case, a clustering method is used to select a representative pattern from each word voice of many speakers.
しかし、標準パターンを作成するには、話者の単語音声
から求めた音声サンプルのデータベースを用い、これに
基づいて標準パターンの作成を行なっておシ、同一話者
による同一単語音声であっても、音声入力手段が例えば
比較的忠実度の高いマイクロホンの場合と忠実度の低い
電話機の場合との如く異れば、変換周波数特性の差異に
より標準パターンも異なるため、入力手段の変更に応じ
その都度音声サンプルを収集し、かつ、データベースを
更新しなければならず、この操作が面倒となる問題を生
じている。However, in order to create a standard pattern, a database of speech samples obtained from the word speech of the speaker is used, and the standard pattern is created based on this. If the audio input means is different, for example, a microphone with relatively high fidelity and a telephone with low fidelity, the standard pattern will also differ due to the difference in conversion frequency characteristics. Voice samples must be collected and the database must be updated, creating a problem in which this operation becomes cumbersome.
前述の問題を解決するため、本発明はつぎの手段によシ
構成するものとなっている。In order to solve the above-mentioned problem, the present invention is constructed by the following means.
すなわち、上述の標準パターンを作成する装置において
、データベースの与えられる標準パターン作成回路の入
力側へ設けられた周波数特性の制御可能なp波器を備え
たものである。That is, the apparatus for creating the standard pattern described above is provided with a p-wave generator whose frequency characteristics can be controlled, which is provided on the input side of the standard pattern creating circuit provided with the database.
したがって、−旦音声サンプルを収集し、データベース
を作成しておけば、入力手段の周波数特性に応じてP波
器の周波数特性を制御することにより、同一のデータベ
ースに基づき、入力手段の特性と対応した特性の標準パ
ターンを作成することができる。Therefore, once voice samples are collected and a database is created, by controlling the frequency characteristics of the P-wave device according to the frequency characteristics of the input means, it is possible to correspond to the characteristics of the input means based on the same database. It is possible to create standard patterns with specific characteristics.
以下、実施例を示す図によって本発明の詳細な説明する
。Hereinafter, the present invention will be explained in detail with reference to figures showing examples.
第1図はブロック図であシ、磁気テープ等の記憶装置に
より記録された音声サンプルのデータベース1は、ディ
ジタルフィルタ等の周波数特性を制御することの可能な
ろ波器2を介し、標準パターン作成回路3へ与えられ、
同回路3の出力として音声認識用の標準パターン4が作
成され、メモリ等へ格納されるものとなっている。FIG. 1 is a block diagram. A database 1 of audio samples recorded on a storage device such as a magnetic tape is processed through a standard pattern creation circuit through a filter 2 that can control frequency characteristics such as a digital filter. given to 3,
A standard pattern 4 for voice recognition is created as an output of the circuit 3 and is stored in a memory or the like.
ここにおいて、データベース1は、比較的高忠実度のマ
イクロホン等に工9収集された音声サンプルから作成さ
れておシ、音声認識を行なう音声が同等の入力手段によ
り与えられる場合は、第2図に周波数f対振幅大の特性
を示すとおり、平坦特性Fがp波器2の周波数特性とし
て制御により設定され、入力手段が電話機の場合には、
これの周波数特性に応じ同等の特性Tが同様に設定され
る。Here, the database 1 is created from voice samples collected using a relatively high-fidelity microphone, etc. If the voice for voice recognition is given by an equivalent input means, the database 1 is shown in FIG. As shown in the characteristic of frequency f vs. large amplitude, when the flat characteristic F is set by control as the frequency characteristic of the p-wave device 2, and the input means is a telephone,
An equivalent characteristic T is similarly set according to this frequency characteristic.
したがって、特性Fの設定により標準パターン4は、マ
イクロホンを介する入力音声のパターンと等価なものと
なシ、特性Tの設定によっては、標準パターン4が電話
機を介する入力音声のパターンと等価なものとなるため
、同一のデータベース1を用いながら、入力手段の変更
に応じた標準パターン4を容易に作成することができる
。Therefore, depending on the setting of characteristic F, standard pattern 4 is equivalent to the pattern of input audio through a microphone, and depending on the setting of characteristic T, standard pattern 4 is equivalent to the pattern of input audio through a telephone. Therefore, while using the same database 1, it is possible to easily create a standard pattern 4 that corresponds to a change in the input means.
なお、p波器2は、標準パターン作成回路3の入力側へ
設ければよく、中間に他の回路が介在しても同様であり
、P波器2の特性は第2図のもののみならず、入力手段
の特性に応じて設定すればよい。Note that the P-wave device 2 may be provided on the input side of the standard pattern creation circuit 3, and the same applies even if another circuit is interposed in between.The characteristics of the P-wave device 2 are those shown in FIG. First, it may be set according to the characteristics of the input means.
以上の説明により明らかなとお9本発明によれば、標準
パターン作成回路の入力側へ周波数特性の制御可能なp
波器を設け、これを介して音声サンプルのデータベース
を与えるものとしたため、同一のデータベースから入力
手段に応じた標準パターンを作成することが自在となり
、音声サンプル収集の手間が省略できるため、音声認識
用の標準パターン作成において顕著な効果が得られる。As is clear from the above explanation, according to the present invention, a controllable frequency characteristic pixel is provided on the input side of the standard pattern creation circuit.
By providing a voice sample database through which a voice sample database is provided, it is possible to create a standard pattern according to the input method from the same database, and the trouble of voice sample collection can be omitted, making it possible to improve voice recognition. A remarkable effect can be obtained in the creation of standard patterns for applications.
図は本発明の実施例を示し、第1図はブロック図、第2
図はp波器の周波数特性を示す図である。
1・・・・データベース、2・・・・P波器、3・・・
・標準パターン作成回路、4・・・・標準パターン。The figures show embodiments of the present invention, with Figure 1 being a block diagram and Figure 2 being a block diagram.
The figure shows the frequency characteristics of a p-wave device. 1...Database, 2...P-wave device, 3...
・Standard pattern creation circuit, 4...Standard pattern.
Claims (1)
パターンを作成する装置において、前記データベースの
与えられる標準パターン作成回路の入力側へ設けられた
周波数特性の制御可能なろ波器を備えたことを特徴とす
る音声認識用標準パターン作成装置。A device for creating a standard pattern for speech recognition based on a database of speech samples, characterized by comprising a filter whose frequency characteristics can be controlled and provided on the input side of a standard pattern creation circuit provided with the database. Standard pattern creation device for speech recognition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP62254731A JPH0199093A (en) | 1987-10-12 | 1987-10-12 | Reference pattern generator for voice recognition |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP62254731A JPH0199093A (en) | 1987-10-12 | 1987-10-12 | Reference pattern generator for voice recognition |
Publications (1)
Publication Number | Publication Date |
---|---|
JPH0199093A true JPH0199093A (en) | 1989-04-17 |
Family
ID=17269071
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
JP62254731A Pending JPH0199093A (en) | 1987-10-12 | 1987-10-12 | Reference pattern generator for voice recognition |
Country Status (1)
Country | Link |
---|---|
JP (1) | JPH0199093A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002278590A (en) * | 2001-03-15 | 2002-09-27 | Ricoh Co Ltd | Speech recognition model generation device, method for generating speech recognition model, speech recognition device, speech recognition method, speech recognition system and recording medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS57115600A (en) * | 1981-01-08 | 1982-07-19 | Sanyo Electric Co | Voice recognition apparatus |
JPS59132000A (en) * | 1983-01-19 | 1984-07-28 | 松下電器産業株式会社 | Preparation of standard voice pattern |
JPS62124599A (en) * | 1985-11-26 | 1987-06-05 | 株式会社東芝 | Voice recognition equipment |
-
1987
- 1987-10-12 JP JP62254731A patent/JPH0199093A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS57115600A (en) * | 1981-01-08 | 1982-07-19 | Sanyo Electric Co | Voice recognition apparatus |
JPS59132000A (en) * | 1983-01-19 | 1984-07-28 | 松下電器産業株式会社 | Preparation of standard voice pattern |
JPS62124599A (en) * | 1985-11-26 | 1987-06-05 | 株式会社東芝 | Voice recognition equipment |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002278590A (en) * | 2001-03-15 | 2002-09-27 | Ricoh Co Ltd | Speech recognition model generation device, method for generating speech recognition model, speech recognition device, speech recognition method, speech recognition system and recording medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP1741313B1 (en) | A method and system for sound source separation | |
KR890017996A (en) | Fitting device for artificial auditory organ using vector and method | |
CN101883304A (en) | Compensation system and method for sound reproduction | |
CN1151077A (en) | Method for reproducing audio signals and apparatus therefor | |
JP2791036B2 (en) | Audio processing device | |
Reiss et al. | Applications of cross-adaptive audio effects: Automatic mixing, live performance and everything in between | |
CN1552171A (en) | Audio reproducing device | |
JPH0199093A (en) | Reference pattern generator for voice recognition | |
DE112009005147T5 (en) | System and method for modifying an audio signal | |
US5893068A (en) | Method of expanding a frequency range of a digital audio signal without increasing a sampling rate | |
JP4185984B2 (en) | Sound signal processing apparatus and processing method | |
JPH06289898A (en) | Speech signal processor | |
JPS63149699A (en) | Voice input/output device | |
JPH06250695A (en) | Method and device for pitch control | |
Fletcher | Stereophonic reproduction from film | |
US1807940A (en) | Sound control apparatus | |
Olson | Trends in Sound Reproduction Research | |
EP0630108A2 (en) | A method of expanding the frequency range of a digital audio signal | |
JPS6367400B2 (en) | ||
Moftah et al. | Language recognition from distorted speech: Comparison of techniques | |
JPS6287994A (en) | Voice recognition dictionary updating system | |
Thienhaus | Principal Considerations on the Artistic Qualities of Musical Sound | |
Brandtsegg et al. | Applications of Cross-Adaptive Audio Effects: Automatic Mixing, Live Performance and Everything in Between | |
Sherman | Binaural sound reproduction at home | |
JPS63147200A (en) | Voice parameter correction system |